|
Activity Number:
|
533
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Thursday, August 2, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
| Abstract - #308746 |
|
Title:
|
Bandwidth Selection for Weighted Kernel Density Estimation
|
|
Author(s):
|
Bin Wang*+
|
|
Companies:
|
University of South Alabama
|
|
Address:
|
ILB 325 University Blvd, Mobile, AL, 36688,
|
|
Keywords:
|
informative censoring ; nonparametric ; bandwidth selection ; weighted kernel density estimate ; missing data
|
|
Abstract:
|
Censoring is common in survival data. For noninformatively censored data, a weighted kernel density estimator can be used to estimate the density of the survival times. However, bandwidth selection could be a problem by reweighting the kernels. In this paper, we proposed to justify the bias cause by noninformative/informative censoring by a weighted kernel method and the bandwidth selection will be discussed. The performance of the new method will be illustrated via Monte Carlo studies.
|
